Time Series Momentum: The Strategy Behind Billions in Hedge Fund Returns

By Prahlad Menon 7 min read

In 2012, a trio of researchers — Tobias Moskowitz (University of Chicago Booth), Yao Hua Ooi (AQR Capital Management), and Lasse Pedersen (NYU Stern) — published a 23-page paper in the Journal of Financial Economics that quietly became one of the most cited and traded ideas in modern finance.

The paper was called “Time Series Momentum.”

It documented something hedge funds had been doing in private for decades — and put it on academic record for the first time.

What Is Time Series Momentum?

The core idea is deceptively simple: an asset that has been going up tends to keep going up. An asset that has been going down tends to keep going down.

More precisely: if an asset’s return over the past 12 months has been positive, you go long. If it’s been negative, you go short. You hold for one month. Then you reassess.

That’s it. No earnings analysis. No CEO interviews. No macro forecasts. Just: has this thing been trending?

This is distinct from the more commonly discussed cross-sectional momentum — which says “buy the winners and sell the losers relative to each other.” Time series momentum doesn’t care about relative performance. It asks: has this asset beaten itself? Has it been positive over its own past history?

The difference matters. An asset can be the best performer in a bad market and still have negative time series momentum. Time series momentum is absolute, not relative.

What the Paper Found

Moskowitz, Ooi, and Pedersen studied 58 liquid futures contracts across four asset classes — equity indices, fixed income, commodities, and currencies — spanning 25 years of data (1985–2009).

The findings were striking:

It works everywhere. Time series momentum produced positive risk-adjusted returns in every single asset class. Equities, bonds, commodities, currencies — the effect was consistent and persistent across all of them.

The 12-month lookback window is the sweet spot. Shorter windows (1–3 months) are noisy. Longer windows (beyond 12 months) start to revert. The 12-month signal captures the trend without chasing it into its own reversal.

It’s not explained by known risk factors. When the researchers ran the strategy’s returns against the standard Fama-French factors (size, value, market beta) and cross-sectional momentum, a large unexplained alpha remained. This isn’t just a repackaging of something we already knew.

It performs best in crisis. This is the finding that made hedge fund managers pay attention. During the worst equity market drawdowns — the 2008 financial crisis, the dot-com bust, the early 1990s recession — time series momentum delivered its strongest returns. It’s naturally long volatility. When markets trend hard in one direction (as they do in crises), the strategy is already positioned.

The annualized Sharpe ratio of a diversified time series momentum portfolio across all 58 markets: approximately 1.0 over 25 years. That’s exceptional for a fully systematic, rule-based strategy with no optimization.

Why Does It Work?

The paper offers two complementary explanations — one behavioral, one structural.

The behavioral story: Markets underreact to new information, then overreact. When a piece of positive news hits, investors don’t immediately price in the full impact. They update gradually, anchored to past prices. This slow adjustment creates the trending behavior that time series momentum exploits. Eventually, the trend overshoots (overreaction), which is why longer lookback windows start to reverse.

The structural story: Hedgers systematically take the other side. Commodity producers hedge by selling futures. Airlines hedge by buying oil futures. These hedgers are willing to pay a premium for certainty — which means they systematically lose to speculators who absorb that risk. Time series momentum is, in part, a risk premium paid by hedgers to speculators. This is why it persists: it’s not a fluke; it’s a structural transfer.

Both explanations suggest the effect won’t simply disappear once it becomes known — it’s rooted in human psychology and institutional incentives that don’t change easily.

How Hedge Funds Use It

The strategy described in the paper maps almost directly onto what’s called Managed Futures or CTA (Commodity Trading Advisor) funds. These are some of the largest systematic hedge funds in the world — firms like AQR, Winton, Man AHL, and Millburn Ridgefield.

The basic implementation:

  1. Universe: Liquid futures across equity indices, rates, commodities, FX — typically 50–150 instruments
  2. Signal: Sign of the 12-month return (or a weighted combination of 1, 3, 6, and 12-month signals)
  3. Sizing: Position size proportional to signal strength, scaled by recent volatility (so you take the same risk per trade, not the same dollar amount)
  4. Rebalancing: Monthly or continuous

Volatility scaling is crucial. A 10% move in crude oil is not the same risk as a 10% move in 10-year Treasuries. Normalizing by volatility ensures the portfolio doesn’t become inadvertently dominated by the most volatile markets.

The result is a portfolio that is long assets in uptrends and short assets in downtrends, constantly rebalancing as trends emerge and fade.

The Crisis Alpha Property

The most commercially important finding in the paper is what practitioners call crisis alpha.

During the 10 worst months for the S&P 500 between 1985 and 2009, a diversified time series momentum portfolio returned an average of +7.4%. During normal months, it earned modest but consistent positive returns.

This negative correlation with equity drawdowns makes time series momentum a rare beast: a strategy that improves the risk-adjusted return of a traditional 60/40 portfolio when you add it as an allocation. Most alternative strategies promise diversification but deliver it only in calm markets. Managed futures actually delivers it when you need it.

This is why pension funds and endowments — which need to survive long drawdowns — have become major allocators to CTA strategies.

The Limitations Worth Knowing

No strategy survives indefinitely without caveats.

It struggles in range-bound markets. When assets chop without trending — as many did in 2011–2012 and again in 2018 — time series momentum generates false signals and incurs transaction costs without meaningful return. CTAs had a difficult decade in the 2010s partially for this reason.

Capacity constraints at scale. The strategy works beautifully on futures, which are liquid and cheap to trade. Applying it to individual stocks or less liquid instruments introduces friction that can erode the edge.

The 2010s anomaly. From roughly 2012 to 2021, extraordinary central bank intervention suppressed volatility and created unusually mean-reverting markets. The strategy underperformed its historical record during this period — though it staged a strong comeback in 2022 when inflation caused clear, sustained trends across asset classes.

It’s now widely known. The paper has been cited thousands of times. Every quant fund has read it. Whether broad adoption has arbitraged away some of the edge is debated — but the structural explanations (hedger premiums, behavioral anchoring) suggest meaningful persistence.

The Bottom Line

Time series momentum is, at its core, a disciplined answer to a simple question: Is this thing trending?

The Moskowitz-Ooi-Pedersen paper didn’t invent the idea — trend-following traders have existed for over a century — but it provided the first rigorous academic confirmation that the effect is real, persistent, and not explained by existing risk models.

For anyone building systematic strategies, managing a portfolio through downturns, or trying to understand what the best macro hedge funds actually do: this 23-page paper is required reading.

You can find the full paper at AQR’s research library — free to download.

If this got you thinking about systematic trading and market signals, here’s more from our corner of the internet:

On themenonlab.com:

On Ray’s Finance Blog (signals.themenonlab.com):

The gap between academic theory and live market application is where most strategies live or die. The Moskowitz paper tells you what works. The daily brief tells you what’s happening right now.

Past performance in academic backtests does not guarantee future results. Time series momentum, like all systematic strategies, involves real risk of loss and requires disciplined implementation.